Re: [boost] [Review][PQS] Physical Quantities

From: Andy Little (andy_at_[hidden])
Date: 2006-06-05 10:57:46

"Jarrad Waterloo"  wrote:

> Wonderful library! I hope it becomes part of boost.

Thanks.  (Maybe ... if  it gets some more  reviews  ...  (and positive) in the 
home stretch ... ).

> Even if we get 't1_?',
> how committed are you to providing 't2_?' and 't3_?'?

t2 was in a previous version and it is useful, should go in. t3 would be a
clumsy great beast, but would be useful for retrieving quantities from XML or
passed as distributed objects:

http://unitsml.nist.gov/

>Further, this may or
> not be the right place for this and I know you are big on the SI or similar
> units but has there been any consideration to adding monetary types. The
> requirements would be a monetary base class with specializations for
> dollars, pounds, euros and expandable to any type of currency. The further
> needs to be a global, static or facet, that is replaceable and stores the
> conversion factors that way it could have a default text file implementation
> that could be replaced with an database implementation.

One reason the PQS cant handle money units is that their relationship isnt
constant or even necessarily agreed upon by any two different authorities at one
time. Physical quantities have nominally constant conversion factors. That
impacts in the way the library can do a lot of computations at compile time. I
know this issue comes up a lot. Why is money not a physical quantity. I dont
really have an answer, except to say that I the SI  dont consider money to be a
physical quantity.

There must also be a need for a type that works out (say) the cost of applying
paint to a certain area. Its another field that PQS could be expanded into. The
scope of that library is potentially very wide of course!

regards
Andy Little